A method for capturing the tacit knowledge in the surface finishing skill by demonstration for programming a robot

Abstract

This paper describes a methodology for capturing the tacit knowledge of the manual grinding and polishing process. Key Process Variables (KPVs) i.e. contact force, tool path, feed rate, etc. of the manual operator performing the task are captured with a `sensorised' hand-held belt grinder, while the changes to the work-piece geometry is captured using a 3-D laser scanner. These KPVs are fed into an analytical material removal model to generate a material removal profile, which can then be calibrated using the actual material removed determined from the manual surface finishing process. The skill of the surface finishing skill is encapsulated in this material removal model and reduces the need for costly robotic Design of Experiment (DoE) trials with test coupons to develop empirical material removal models. Parts from the production process require different processing variables, but the common objective is to generate various material removal maps in order to manufacture a part of a desired form and dimension. The metal removal rate (MRR) model could then be utilized by the industrial robot to determine suitable polishing parameters to accomplish the polishing task. The characteristics of the skilled worker's captured motions can then be extracted and used for optimization of the industrial robot polishing tool path.